Peter Korsgaard [Wed, 11 Aug 2010 01:01:24 +0000 (18:01 -0700)]
scsi: use __uX types for headers exported to user space
Commit 9e4f5e29 ("FC Pass Thru support") exported a number of header files
in include/scsi to user space, but didn't change the uX types to the
userspace-compatible __uX types. Without that you'll get compile errors
when including them - E.G.:
include/scsi/scsi.h:145: error: expected specifier-qualifier-list before `u8'
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk> Cc: Boaz Harrosh <bharrosh@panasas.com> Cc: James Smart <james.smart@emulex.com> Cc: James Bottomley <James.Bottomley@HansenPartnership.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
(cherry picked from commit 083c8c1e60e5c27a277e87dbeb6b89b47937559f) Cc: Thomas Bork <tom@eisfair.net> Signed-off-by: Willy Tarreau <w@1wt.eu>
Sometimes, warnings about ioctls to partition happen often enough that they
form majority of the warnings in the kernel log and users complain. In some
cases warnings are about ioctls such as SG_IO so it's not good to get rid of
the warnings completely as they can ease debugging of userspace problems
when ioctl is refused.
Since I have seen warnings from lots of commands, including some proprietary
userspace applications, I don't think disallowing the ioctls for processes
with CAP_SYS_RAWIO will happen in the near future if ever. So lets just
stop warning for processes with CAP_SYS_RAWIO for which ioctl is allowed.
CC: Paolo Bonzini <pbonzini@redhat.com> CC: James Bottomley <JBottomley@parallels.com> CC: linux-scsi@vger.kernel.org Acked-by: Paolo Bonzini <pbonzini@redhat.com> Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@kernel.dk>
[bwh: Backported to 3.2: use ENOTTY, not ENOIOCTLCMD] Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit f45c9a6eec20cd712421c442785e7a4e9215a230) Cc: Thomas Bork <tom@eisfair.net> Signed-off-by: Willy Tarreau <w@1wt.eu>
Don't zero out bits 15..12 of the data value in `das08jr_ao_winsn()` as
that knobbles the upper three-quarters of the output range for the
'das08jr-16-ao' board.
Some low-level comedi drivers (incorrectly) point `dev->read_subdev` or
`dev->write_subdev` to a subdevice that does not support asynchronous
commands. Comedi's poll(), read() and write() file operation handlers
assume these subdevices do support asynchronous commands. In
particular, they assume `s->async` is valid (where `s` points to the
read or write subdevice), which it won't be if it has been set
incorrectly. This can lead to a NULL pointer dereference.
Check `s->async` is non-NULL in `comedi_poll()`, `comedi_read()` and
`comedi_write()` to avoid the bug.
`parse_insn()` is dereferencing the user-space pointer `insn->data`
directly when handling the `INSN_INTTRIG` comedi instruction. It
shouldn't be using `insn->data` at all; it should be using the separate
`data` pointer passed to the function. Fix it.
Correct a direct dereference of I/O memory to use an appropriate I/O
memory access function. Note that the pointer being dereferenced is not
currently tagged with `__iomem` but I plan to correct that for 3.7.
`s626_enc_insn_config()` is incorrectly dereferencing `insn->data` which
is a pointer to user memory. It should be dereferencing the separate
`data` parameter that points to a copy of the data in kernel memory.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
When `do_cmd_ioctl()` allocates memory for the kernel copy of a channel
list, it frees any previously allocated channel list in
`async->cmd.chanlist` and replaces it with the new one. However, if the
device is ever removed (or "detached") the cleanup code in
`cleanup_device()` in "drivers.c" does not free this memory so it is
lost.
A sensible place to free the kernel copy of the channel list is in
`do_become_nonbusy()` as at that point the comedi asynchronous command
associated with the channel list is no longer valid. Free the channel
list in `do_become_nonbusy()` instead of `do_cmd_ioctl()` and clear the
pointer to prevent it being freed more than once.
Note that `cleanup_device()` could be called at an inappropriate time
while the comedi device is open, but that's a separate bug not related
to this this patch.
Éric Piel reported a kernel oops in the "comedi_test" module. It was a
NULL pointer dereference within `waveform_ai_interrupt()` (actually a
timer function) that sometimes occurred when a running asynchronous
command is cancelled (either by the `COMEDI_CANCEL` ioctl or by closing
the device file).
This seems to be a race between the caller of `waveform_ai_cancel()`
which on return from that function goes and tears down the running
command, and the timer function which uses the command. In particular,
`async->cmd.chanlist` gets freed (and the pointer set to NULL) by
`do_become_nonbusy()` in "comedi_fops.c" but a previously scheduled
`waveform_ai_interrupt()` timer function will dereference that pointer
regardless, leading to the oops.
Fix it by replacing the `del_timer()` call in `waveform_ai_cancel()`
with `del_timer_sync()`.
Tuomas <tvainikk _at_ gmail _dot_ com> reported problems getting
meaningful output from a Lab-PC+ in differential mode for AI cmds, but
AI insn reads gave correct readings. He tracked it down to two
problems, one of which is addressed by this patch.
It seems that writing to the command3 register after writing to the
command4 register in `labpc_ai_cmd()` messes up the differential
reference bit setting in the command4 register. Set up the command4
register after the command3 register (as in `labpc_ai_rinsn()`) to avoid
the problem.
Thanks to Tuomas for suggesting the fix.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
Tuomas <tvainikk _at_ gmail _dot_ com> reported problems getting
meaningful output from a Lab-PC+ in differential mode for AI cmds, but
AI insn reads gave correct readings. He tracked it down to two
problems, one of which is addressed by this patch.
It seems the setting of the channel bits for particular scanning modes
was incorrect for differential mode. (Only half the number of channels
are available in differential mode; comedi refers to them as channels 0,
1, 2 and 3, but the hardware documentation refers to them as channels 0,
2, 4 and 6.) In differential mode, the setting of the channel enable
bits in the command1 register should depend on whether the scan enable
bit is set. Effectively, we need to double the comedi channel number
when the scan enable bit is not set in differential mode. The scan
enable bit gets set when the AI scan mode is `MODE_MULT_CHAN_UP` or
`MODE_MULT_CHAN_DOWN`, and gets cleared when the AI scan mode is
`MODE_SINGLE_CHAN` or `MODE_SINGLE_CHAN_INTERVAL`. The existing test
for whether the comedi channel number needs to be doubled in
differential mode is incorrect in `labpc_ai_cmd()`. This patch corrects
the test.
Thanks to Tuomas for suggesting the fix.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
Good job. I was too lazy to bisect for bad commit;)
Reading the code I found problematic kthread_should_stop call from netlink
connector which causes the oops. After applying a patch, I've been testing
owfs+w1 setup for nearly two days and it seems to work very reliable (no
hangs, no memleaks etc).
More detailed description and possible fix is given below:
Function w1_search can be called from either kthread or netlink callback.
While the former works fine, the latter causes oops due to kthread_should_stop
invocation.
This patch adds a check if w1_search is serving netlink command, skipping
kthread_should_stop invocation if so.
Jiri Slaby [Sun, 5 Jun 2011 20:51:49 +0000 (22:51 +0200)]
serial: 8250, increase PASS_LIMIT
With virtual machines like qemu, it's pretty common to see "too much
work for irq4" messages nowadays. This happens when a bunch of output
is printed on the emulated serial console. This is caused by too low
PASS_LIMIT. When ISR loops more than the limit, it spits the message.
I've been using a kernel with doubled the limit and I couldn't see no
problems. Maybe it's time to get rid of the message now?
[2.6.32: background info from Ram Gupta]
> I need a patch for serial driver that increases PASS_LIMIT merged in
> 3.1. I am using 2.6.32 kernel which experiences kernel panic
> occasionally. It will be great if you can backport to 2.6.32 and 3.0
> kernel. The commit ID is e7328ae1 serial: 8250, increase PASS_LIMIT
(in acpi_processor_power_init) and never sets the per_cpu idle
device. So when acpi_processor_hotplug on CPU online notification
tries to reference said device it crashes:
When running on 32bit the mce handler could misinterpret
vm86 mode as ring 0. This can affect whether it does recovery
or not; it was possible to panic when recovery was actually
possible.
Fix this by always forcing vm86 to look like ring 3.
[ Backport to 3.0 notes:
Things changed there slightly:
- move mce_get_rip() up. It fills up m->cs and m->ip values which
are evaluated in mce_severity(). Therefore move it up right before
the mce_severity call. This seem to be another bug in 3.0?
- Place the backport (fix m->cs in V86 case) to where m->cs gets
filled which is mce_get_rip() in 3.0
]
Signed-off-by: Andi Kleen <ak@linux.intel.com> Signed-off-by: Tony Luck <tony.luck@intel.com> Signed-off-by: Thomas Renninger <trenn@suse.de> Reviewed-by: Tony Luck <tony.luck@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
On hosts without the XSAVE support unprivileged local user can trigger
oops similar to the one below by setting X86_CR4_OSXSAVE bit in guest
cr4 register using KVM_SET_SREGS ioctl and later issuing KVM_RUN
ioctl.
QEMU first retrieves the supported features via KVM_GET_SUPPORTED_CPUID
and then sets them later. So guest's X86_FEATURE_XSAVE should be masked
out on hosts without X86_FEATURE_XSAVE, making kvm_set_cr4 with
X86_CR4_OSXSAVE fail. Userspaces that allow specifying guest cpuid with
X86_FEATURE_XSAVE even on hosts that do not support it, might be
susceptible to this attack from inside the guest as well.
Allow setting X86_CR4_OSXSAVE bit only if host has XSAVE support.
Signed-off-by: Petr Matousek <pmatouse@redhat.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
[bwh: Backported to 2.6.32: XSAVE is not supported at all, so always
deny setting OSXSAVE] Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
If the guest specifies a IOAPIC_REG_SELECT with an invalid value and follows
that with a read of the IOAPIC_REG_WINDOW KVM does not properly validate
that request. ioapic_read_indirect contains an
ASSERT(redir_index < IOAPIC_NUM_PINS), but the ASSERT has no effect in
non-debug builds. In recent kernels this allows a guest to cause a kernel
oops by reading invalid memory. In older kernels (pre-3.3) this allows a
guest to read from large ranges of host memory.
Tested: tested against apic unit tests.
Signed-off-by: Andrew Honig <ahonig@google.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com> Cc: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
RHEL5 i386 guests register non 32-byte aligned addresses:
kvm-clock: cpu 1, msr 0:3018aa5, secondary cpu clock
kvm-clock: cpu 2, msr 0:301f8e9, secondary cpu clock
kvm-clock: cpu 3, msr 0:302672d, secondary cpu clock
Check for an address+len that would cross page boundary
instead.
If the guest sets the GPA of the time_page so that the request to update the
time straddles a page then KVM will write onto an incorrect page. The
write is done byusing kmap atomic to get a pointer to the page for the time
structure and then performing a memcpy to that page starting at an offset
that the guest controls. Well behaved guests always provide a 32-byte aligned
address, however a malicious guest could use this to corrupt host kernel
memory.
Tested: Tested against kvmclock unit test.
Signed-off-by: Andrew Honig <ahonig@google.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com> Cc: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
The hypervisor will trap it. However without this patch,
we would crash as the .read_tscp is set to NULL. This patch
fixes it and sets it to the native_read_tscp call.
In paravirtualized x86_64 kernels, vmalloc_fault may cause an oops
when lazy MMU updates are enabled, because set_pgd effects are being
deferred.
One instance of this problem is during process mm cleanup with memory
cgroups enabled. The chain of events is as follows:
- zap_pte_range enables lazy MMU updates
- zap_pte_range eventually calls mem_cgroup_charge_statistics,
which accesses the vmalloc'd mem_cgroup per-cpu stat area
- vmalloc_fault is triggered which tries to sync the corresponding
PGD entry with set_pgd, but the update is deferred
- vmalloc_fault oopses due to a mismatch in the PUD entries
Investigation determined that the bug triggered when reading
system RAM at the 4G mark. On this system, that was the first
address using 1G pages for the virt->phys direct mapping so the
PUD is pointing to a physical address, not a PMD page.
The problem is that the page table walker in kern_addr_valid() is
not checking pud_large() and treats the physical address as if
it was a PMD. If it happens to look like pmd_none then it'll
silently fail, probably returning zeros instead of real data. If
the data happens to look like a present PMD though, it will be
walked resulting in the oops above.
This patch adds the necessary pud_large() check.
Unfortunately the problem was not readily reproducible and now
they are running the backup program without accessing
/proc/kcore so the patch has not been validated but I think it
makes sense.
At the moment the MSR driver only relies upon file system
checks. This means that anything as root with any capability set
can write to MSRs. Historically that wasn't very interesting but
on modern processors the MSRs are such that writing to them
provides several ways to execute arbitary code in kernel space.
Sample code and documentation on doing this is circulating and
MSR attacks are used on Windows 64bit rootkits already.
In the Linux case you still need to be able to open the device
file so the impact is fairly limited and reduces the security of
some capability and security model based systems down towards
that of a generic "root owns the box" setup.
Therefore they should require CAP_SYS_RAWIO to prevent an
elevation of capabilities. The impact of this is fairly minimal
on most setups because they don't have heavy use of
capabilities. Those using SELinux, SMACK or AppArmor rules might
want to consider if their rulesets on the MSR driver could be
tighter.
Signed-off-by: Alan Cox <alan@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Horses <stable@kernel.org> Signed-off-by: Ingo Molnar <mingo@kernel.org>
[dannf: backported to Debian's 2.6.32] Signed-off-by: Willy Tarreau <w@1wt.eu>
Drew Jones while working on CVE-2013-0190 found that that unprivileged guest user
in 32bit PV guest can use to crash the > guest with the panic like this:
Petr says: "
I've analysed the bug and I think that xen_iret() cannot cope with
mangled DS, in this case zeroed out (null selector/descriptor) by either
xen_failsafe_callback() or RESTORE_REGS because the corresponding LDT
entry was invalidated by the reproducer. "
Jan took a look at the preliminary patch and came up a fix that solves
this problem:
"This code gets called after all registers other than those handled by
IRET got already restored, hence a null selector in %ds or a non-null
one that got loaded from a code or read-only data descriptor would
cause a kernel mode fault (with the potential of crashing the kernel
as a whole, if panic_on_oops is set)."
The way to fix this is to realize that the we can only relay on the
registers that IRET restores. The two that are guaranteed are the
%cs and %ss as they are always fixed GDT selectors. Also they are
inaccessible from user mode - so they cannot be altered. This is
the approach taken in this patch.
Another alternative option suggested by Jan would be to relay on
the subtle realization that using the %ebp or %esp relative references uses
the %ss segment. In which case we could switch from using %eax to %ebp and
would not need the %ss over-rides. That would also require one extra
instruction to compensate for the one place where the register is used
as scaled index. However Andrew pointed out that is too subtle and if
further work was to be done in this code-path it could escape folks attention
and lead to accidents.
Reviewed-by: Petr Matousek <pmatouse@redhat.com> Reported-by: Petr Matousek <pmatouse@redhat.com> Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com> Signed-off-by: Jan Beulich <jbeulich@suse.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
[dannf: backported to Debian's 2.6.32] Signed-off-by: Willy Tarreau <w@1wt.eu>
Romain Francoise [Wed, 23 Jan 2013 10:58:36 +0000 (11:58 +0100)]
x86, random: make ARCH_RANDOM prompt if EMBEDDED, not EXPERT
Before v2.6.38 CONFIG_EXPERT was known as CONFIG_EMBEDDED but the
Kconfig entry was not changed to match when upstream commit 628c6246d47b85f5357298601df2444d7f4dd3fd ("x86, random: Architectural
inlines to get random integers with RDRAND") was backported.
Matthew Garrett [Wed, 6 Jul 2011 20:52:37 +0000 (16:52 -0400)]
x86: Don't use the EFI reboot method by default
Testing suggests that at least some Lenovos and some Intels will
fail to reboot via EFI, attempting to jump to an unmapped
physical address. In the long run we could handle this by
providing a page table with a 1:1 mapping of physical addresses,
but for now it's probably just easier to assume that ACPI or
legacy methods will be present and reboot via those.
[2.6.32: additional background information from Jonathan below]
>
> Please consider
>
> f70e957cda22 x86: Don't use the EFI reboot method by default,
> 2011-07-06
>
> for application to the 2.6.32.y and 2.6.34.y trees. The patch was
> applied upstream late in the 3.0 cycle, so newer kernels don't need
> it.
>
> In 2011, Keith Ward wrote[1]:
>
> > When attempting to reboot my my UEFI enabled system, the system hangs when
> > calling reboot requiring me to manually reset the system via the reset switch.
> >
> > Screenshot: http://twitgoo.com/29bq1c
>
> Ben Hutchings writes[1]:
>
> > Version: 3.0.0-1
> >
> > I also had this problem on my own system, but it is fixed now.
> > I bisected the fix to:
> >
> > commit f70e957cda22d309c769805cbb932407a5232219
> > Author: Matthew Garrett <mjg@redhat.com>
> > Date: Wed Jul 6 16:52:37 2011 -0400
> >
> > x86: Don't use the EFI reboot method by default
> >
> > which is basically equivalent to the workaround!
> >
> > I'll also apply this fix to squeeze as it's so simple.
>
> Keith Ward also wrote[1]:
>
> > It seems as if this has recently been reported at Ubuntu's Launchpad as well:
> > https://bugs.launchpad.net/ubuntu/+source/linux/+bug/721576
>
> There are a variety of reports of the same panic at that bug on
> 2.6.32.y-, 2.6.38.y-, and 2.6.39-based kernels. Passing "reboot=a,w"
> on the kernel command line avoids trouble for reporters.
Signed-off-by: Matthew Garrett <mjg@redhat.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Alan Cox <alan@linux.intel.com> Link: http://lkml.kernel.org/r/1309985557-15350-1-git-send-email-mjg@redhat.com Signed-off-by: Ingo Molnar <mingo@elte.hu>
(cherry picked from commit f70e957cda22d309c769805cbb932407a5232219) Cc: Jonathan Nieder <jrnieder@gmail.com> Cc: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
Suresh Siddha [Wed, 24 Oct 2012 22:24:30 +0000 (15:24 -0700)]
x86, ioapic: initialize nr_ioapic_registers early in mp_register_ioapic()
Lin Bao reported that one of the HP platforms failed to boot
2.6.32 kernel, when the BIOS enabled interrupt-remapping and
x2apic before handing over the control to the Linux kernel.
During boot, Linux kernel masks all the interrupt sources
(8259, IO-APIC RTE's), setup the interrupt-remapping hardware
with the OS controlled table and unmasks the 8259 interrupts
but not the IO-APIC RTE's (as the newly setup interrupt-remapping
table and the IO-APIC RTE's are not yet programmed by the kernel).
Shortly after this, IO-APIC RTE's and the interrupt-remapping table
entries are programmed based on the ACPI tables etc. So the
expectation is that any interrupt during this window will be dropped
and not see the intermediate configuration.
In the reported problematic case, BIOS has configured the IO-APIC
in virtual wire-B mode. Between the window of the kernel setting up
new interrupt-remapping table and the IO-APIC RTE's are properly
configured, an interrupt gets routed by the IO-APIC RTE (setup
by the virtual wire-B configuration) and sees the empty
interrupt-remapping table entry, resulting in vt-d fault causing
the platform to generate NMI. And the OS panics on this unexpected NMI.
This problem doesn't happen with more recent kernels and closer
look at the 2.6.32 kernel shows that the code which masks
the IO-APIC RTE's is not working as expected as the nr_ioapic_registers
for each IO-APIC is not yet initialized at this point. In the later
kernels we initialize nr_ioapic_registers much before and
everything works as expected.
For 2.6.[32..34] kernels, fix this issue by initializing
nr_ioapic_registers early in mp_register_ioapic()
[ Relevant upstream commit info:
commit 7716a5c4ff5f1f3dc5e9edcab125cbf7fceef0af
Author: Eric W. Biederman <ebiederm@xmission.com>
Date: Tue Mar 30 01:07:12 2010 -0700
x86, ioapic: Move nr_ioapic_registers calculation to mp_register_ioapic.
As the upstream commit depends on quite a few prior commits
and some followup fixes in the mainline, we just picked
the smallest relevant hunk for fixing the issue at hand.
Problematic platform uses ACPI for IO-APIC, VT-d enumeration etc
and this hunk only touches the ACPI based platforms.
nr_ioapic_reigsters initialization in enable_IO_APIC() is still
retained, so that other configurations like legacy MPS table based
enumeration etc works with no change.
]
snd_seq_timer_open() didn't catch the whole error path but let through
if the timer id is a slave. This may lead to Oops by accessing the
uninitialized pointer.
[2.6.32: additional background from Jonathan below]
> Hi Willy,
>
> Please consider
>
> cfc9b06f0bef ALSA: hda - Add a pin-fix for FSC Amilo Pi1505
>
> for application to the 2.6.32.y tree. Without this patch, the Amilo
> Pi 1505's internal speaker is silent unless a jack is plugged into its
> headphone jack.
>
> Jose Manuel Castroagudin noticed[1] that 2.6.30 is not affected, so
> this seems to be a regression.
>
> The patch was applied upstream during the 2.6.33 merge window, where
> it worked. That said, I didn't manage to track down anyone with a
> Pi1505 to test it against 2.6.32, so thoughts from alsa folks on
> whether this is appropriate for 2.6.32.y would be useful.
>
> Hope that helps,
> Jonathan
>
> [1] http://bugs.debian.org/599582 has many more details.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
(cherry picked from commit cfc9b06f0befe50ef02253f72b76946363549031) Cc: Jonathan Nieder <jrnieder@gmail.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
1. Add more ASUS NB model.
2. Fixed alc663_m51va_setup
M51VA has Digital Mic that NID is 0x12. The record source index is
0x9 for ALC663.
So, to modify the alc663_m51va_setup function to index 0x9
and add analog Mic aupport function alc663_mode1_setup.
3. Add ASUS mode7 and mode8 modules for ALC663
[jn: backport to 2.6.32.y to address http://bugs.debian.org/688564]
Signed-off-by: Kailang Yang <kailang@realtek.com.tw> Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Jonathan Nieder <jrnieder@gmail.com> Tested-by: Bill Allombert <Bill.Allombert@math.u-bordeaux1.fr> # Vaio w/ ALC275 Signed-off-by: Willy Tarreau <w@1wt.eu>
shared_policy_replace() use of sp_alloc() is unsafe. 1) sp_node cannot
be dereferenced if sp->lock is not held and 2) another thread can modify
sp_node between spin_unlock for allocating a new sp node and next
spin_lock. The bug was introduced before 2.6.12-rc2.
Kosaki's original patch for this problem was to allocate an sp node and
policy within shared_policy_replace and initialise it when the lock is
reacquired. I was not keen on this approach because it partially
duplicates sp_alloc(). As the paths were sp->lock is taken are not that
performance critical this patch converts sp->lock to sp->mutex so it can
sleep when calling sp_alloc().
In fuzzing with trinity, lockdep protested "possible irq lock inversion
dependency detected" when isolate_lru_page() reenabled interrupts while
still holding the supposedly irq-safe tree_lock:
isolate_lru_page() is correct to enable interrupts unconditionally:
invalidate_complete_page2() is incorrect to call clear_page_mlock() while
holding tree_lock, which is supposed to nest inside lru_lock.
Both truncate_complete_page() and invalidate_complete_page() call
clear_page_mlock() before taking tree_lock to remove page from radix_tree.
I guess invalidate_complete_page2() preferred to test PageDirty (again)
under tree_lock before committing to the munlock; but since the page has
already been unmapped, its state is already somewhat inconsistent, and no
worse if clear_page_mlock() moved up.
Reported-by: Sasha Levin <levinsasha928@gmail.com> Deciphered-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michel Lespinasse <walken@google.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
In kswapd(), set current->reclaim_state to NULL before returning, as
current->reclaim_state holds reference to variable on kswapd()'s stack.
In rare cases, while returning from kswapd() during memory offlining,
__free_slab() and freepages() can access the dangling pointer of
current->reclaim_state.
Unfortunately with !CONFIG_PAGEFLAGS_EXTENDED, (!PageHead) is false, and
(PageHead) is true, for tail pages. If this is indeed the intended
behavior, which I doubt because it breaks cache cleaning on some ARM
systems, then the nomenclature is highly problematic.
This patch makes sure PageHead is only true for head pages and PageTail
is only true for tail pages, and neither is true for non-compound pages.
[ This buglet seems ancient - seems to have been introduced back in Apr
2008 in commit 6a1e7f777f61: "pageflags: convert to the use of new
macros". And the reason nobody noticed is because the PageHead()
tests are almost all about just sanity-checking, and only used on
pages that are actual page heads. The fact that the old code returned
true for tail pages too was thus not really noticeable. - Linus ]
Signed-off-by: Christoffer Dall <cdall@cs.columbia.edu> Acked-by: Andrea Arcangeli <aarcange@redhat.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Will Deacon <Will.Deacon@arm.com> Cc: Steve Capper <Steve.Capper@arm.com> Cc: Christoph Lameter <cl@linux.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
hugetlb_reserve_pages() can be used for either normal file-backed
hugetlbfs mappings, or MAP_HUGETLB. In the MAP_HUGETLB, semi-anonymous
mode, there is not a VMA around. The new call to resv_map_put() assumed
that there was, and resulted in a NULL pointer dereference:
When called for anonymous (non-shared) mappings, hugetlb_reserve_pages()
does a resv_map_alloc(). It depends on code in hugetlbfs's
vm_ops->close() to release that allocation.
However, in the mmap() failure path, we do a plain unmap_region() without
the remove_vma() which actually calls vm_ops->close().
This is a decent fix. This leak could get reintroduced if new code (say,
after hugetlb_reserve_pages() in hugetlbfs_file_mmap()) decides to return
an error. But, I think it would have to unroll the reservation anyway.
This patch tries to fix a dead loop in async_synchronize_full(), which
could be seen when preemption is disabled on a single cpu machine.
void async_synchronize_full(void)
{
do {
async_synchronize_cookie(next_cookie);
} while (!list_empty(&async_running) || !
list_empty(&async_pending));
}
async_synchronize_cookie() calls async_synchronize_cookie_domain() with
&async_running as the default domain to synchronize.
However, there might be some works in the async_pending list from other
domains. On a single cpu system, without preemption, there is no chance
for the other works to finish, so async_synchronize_full() enters a dead
loop.
It seems async_synchronize_full() wants to synchronize all entries in
all running lists(domains), so maybe we could just check the entry_count
to know whether all works are finished.
Currently, async_synchronize_cookie_domain() expects a non-NULL running
list ( if NULL, there would be NULL pointer dereference ), so maybe a
NULL pointer could be used as an indication for the functions to
synchronize all works in all domains.
Reported-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Signed-off-by: Li Zhong <zhong@linux.vnet.ibm.com> Tested-by: Paul E. McKenney <paulmck@linux.vnet.ibm.com> Tested-by: Christian Kujau <lists@nerdbynature.de> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Dan Williams <dan.j.williams@gmail.com> Cc: Christian Kujau <lists@nerdbynature.de> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Cong Wang <xiyou.wangcong@gmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
cgroup_create_dir() does weird dancing with dentry refcnt. On
success, it gets and then puts it achieving nothing. On failure, it
puts but there isn't no matching get anywhere leading to the following
oops if cgroup_create_file() fails for whatever reason.
A bus_type has a list of devices (klist_devices), but the list and the
subsys_private structure that contains it are not initialized until the
bus_type is registered with bus_register().
The panic/reboot path has fixups that look up devices in pci_bus_type. If
we panic before registering pci_bus_type, the bus_type exists but the list
does not, so mach_reboot_fixups() trips over a null pointer and panics
again:
mach_reboot_fixups
pci_get_device
..
bus_find_device(&pci_bus_type, ...)
bus->p is NULL
Joonsoo reported a problem when panicking before PCI was initialized.
I think this patch should be sufficient to replace the patch he posted
here: https://lkml.org/lkml/2012/12/28/75 ("[PATCH] x86, reboot: skip
reboot_fixups in early boot phase")
Using a recursive call add a non-conflicting region in
__reserve_region_with_split() could result in a stack overflow in the case
that the recursive calls are too deep. Convert the recursive calls to an
iterative loop to avoid the problem.
Tested on a machine containing 135 regions. The kernel no longer panicked
with stack overflow.
Also tested with code arbitrarily adding regions with no conflict,
embedding two consecutive conflicts and embedding two non-consecutive
conflicts.
Signed-off-by: T Makphaibulchoke <tmac@hp.com> Reviewed-by: Ram Pai <linuxram@us.ibm.com> Cc: Paul Gortmaker <paul.gortmaker@gmail.com> Cc: Wei Yang <weiyang@linux.vnet.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Cc: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
The genalloc code uses the bitmap API from include/linux/bitmap.h and
lib/bitmap.c, which is based on long values. Both bitmap_set from
lib/bitmap.c and bitmap_set_ll, which is the lockless version from
genalloc.c, use BITMAP_LAST_WORD_MASK to set the first bits in a long in
the bitmap.
That one uses (1 << bits) - 1, 0b111, if you are setting the first three
bits. This means that the API counts from the least significant bits
(LSB from now on) to the MSB. The LSB in the first long is bit 0, then.
The same works for the lookup functions.
The genalloc code uses longs for the bitmap, as it should. In
include/linux/genalloc.h, struct gen_pool_chunk has unsigned long
bits[0] as its last member. When allocating the struct, genalloc should
reserve enough space for the bitmap. This should be a proper number of
longs that can fit the amount of bits in the bitmap.
However, genalloc allocates an integer number of bytes that fit the
amount of bits, but may not be an integer amount of longs. 9 bytes, for
example, could be allocated for 70 bits.
This is a problem in itself if the Least Significat Bit in a long is in
the byte with the largest address, which happens in Big Endian machines.
This means genalloc is not allocating the byte in which it will try to
set or check for a bit.
This may end up in memory corruption, where genalloc will try to set the
bits it has not allocated. In fact, genalloc may not set these bits
because it may find them already set, because they were not zeroed since
they were not allocated. And that's what causes a BUG when
gen_pool_destroy is called and check for any set bits.
What really happens is that genalloc uses kmalloc_node with __GFP_ZERO
on gen_pool_add_virt. With SLAB and SLUB, this means the whole slab
will be cleared, not only the requested bytes. Since struct
gen_pool_chunk has a size that is a multiple of 8, and slab sizes are
multiples of 8, we get lucky and allocate and clear the right amount of
bytes.
Hower, this is not the case with SLOB or with older code that did memset
after allocating instead of using __GFP_ZERO.
So, a simple module as this (running 3.6.0), will cause a crash when
rmmod'ed.
The function rb_check_pages() was added to make sure the ring buffer's
pages were sane. This check is done when the ring buffer size is modified
as well as when the iterator is released (closing the "trace" file),
as that was considered a non fast path and a good place to do a sanity
check.
The problem is that the check does not have any locks around it.
If one process were to read the trace file, and another were to read
the raw binary file, the check could happen while the reader is reading
the file.
The issues with this is that the check requires to clear the HEAD page
before doing the full check and it restores it afterward. But readers
require the HEAD page to exist before it can read the buffer, otherwise
it gives a nasty warning and disables the buffer.
By adding the reader lock around the check, this keeps the race from
happening.
As kernel_power_off() calls disable_nonboot_cpus(), we may also want to
have kernel_restart() call disable_nonboot_cpus(). Doing so can help
machines that require boot cpu be the last alive cpu during reboot to
survive with kernel restart.
This fixes one reboot issue seen on imx6q (Cortex-A9 Quad). The machine
requires that the restart routine be run on the primary cpu rather than
secondary ones. Otherwise, the secondary core running the restart
routine will fail to come to online after reboot.
In !CORE_DUMP_USE_REGSET case, if elf_note_info_init fails to allocate
memory for info->fields, it frees already allocated stuff and returns
error to its caller, fill_note_info. Which in turn returns error to its
caller, elf_core_dump. Which jumps to cleanup label and calls
free_note_info, which will happily try to free all info->fields again.
BOOM.
Oleg Nesterov [Tue, 19 Feb 2013 17:31:10 +0000 (17:31 +0000)]
wake_up_process() should be never used to wakeup a TASK_STOPPED/TRACED task
wake_up_process() should be never used to wakeup a TASK_STOPPED/TRACED task
CVE-2013-0871
BugLink: http://bugs.launchpad.net/bugs/1129192
wake_up_process() should never wakeup a TASK_STOPPED/TRACED task.
Change it to use TASK_NORMAL and add the WARN_ON().
TASK_ALL has no other users, probably can be killed.
Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Colin King <colin.king@canonical.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
Ben Hutchings [Mon, 26 Nov 2012 03:24:19 +0000 (22:24 -0500)]
signal: Define __ARCH_HAS_SA_RESTORER so we know whether to clear sa_restorer
Vaguely based on upstream commit 574c4866e33d 'consolidate kernel-side
struct sigaction declarations'.
flush_signal_handlers() needs to know whether sigaction::sa_restorer
is defined, not whether SA_RESTORER is defined. Define the
__ARCH_HAS_SA_RESTORER macro to indicate this.
Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Cc: Al Viro <viro@zeniv.linux.org.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
John Johansen [Thu, 21 Mar 2013 12:04:13 +0000 (05:04 -0700)]
ptrace: Fix ptrace when task is in task_is_stopped() state
This patch fixes a regression in ptrace, introduced by commit 9e74eb39
(backport of 9899d11f) which makes assumptions about ptrace behavior
which are not true in the 2.6.32 kernel.
BugLink: http://bugs.launchpad.net/bugs/1145234 9899d11f makes the assumption that task_is_stopped() is not a valid state
in ptrace because it is built on top of a series of patches which change
how the TASK_STOPPED state is tracked (321fb561 which requires d79fdd6d
and several other patches).
Because Lucid does not have the set of patches that make task_is_stopped()
an invalid state in ptrace_check_attach, partially revert 9e74eb39 so
that ptrace_check_attach() correctly handles task_is_stopped(). However
we must replace the assignment of TASK_TRACED with __TASK_TRACED to
ensure TASK_WAKEKILL is cleared.
Signed-off-by: John Johansen <john.johansen@canonical.com> Acked-by: Colin King <colin.king@canonical.com> Acked-by: Stefan Bader <stefan.bader@canonical.com> Acked-by: Luis Henriques <luis.henriques@canonical.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Cc: Oleg Nesterov <oleg@redhat.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
Oleg Nesterov [Tue, 19 Feb 2013 17:31:09 +0000 (17:31 +0000)]
ptrace: ensure arch_ptrace/ptrace_request can never race with SIGKILL
ptrace: ensure arch_ptrace/ptrace_request can never race with SIGKILL
CVE-2013-0871
BugLink: http://bugs.launchpad.net/bugs/1129192
putreg() assumes that the tracee is not running and pt_regs_access() can
safely play with its stack. However a killed tracee can return from
ptrace_stop() to the low-level asm code and do RESTORE_REST, this means
that debugger can actually read/modify the kernel stack until the tracee
does SAVE_REST again.
set_task_blockstep() can race with SIGKILL too and in some sense this
race is even worse, the very fact the tracee can be woken up breaks the
logic.
As Linus suggested we can clear TASK_WAKEKILL around the arch_ptrace()
call, this ensures that nobody can ever wakeup the tracee while the
debugger looks at it. Not only this fixes the mentioned problems, we
can do some cleanups/simplifications in arch_ptrace() paths.
Probably ptrace_unfreeze_traced() needs more callers, for example it
makes sense to make the tracee killable for oom-killer before
access_process_vm().
While at it, add the comment into may_ptrace_stop() to explain why
ptrace_stop() still can't rely on SIGKILL and signal_pending_state().
Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Colin King <colin.king@canonical.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
signal_wake_up(resume => true) is overused. None of ptrace/jctl callers
actually want to wakeup a TASK_WAKEKILL task, but they can't specify the
necessary mask.
Turn signal_wake_up() into signal_wake_up_state(state), reintroduce
signal_wake_up() as a trivial helper, and add ptrace_signal_wake_up()
which adds __TASK_TRACED.
This way ptrace_signal_wake_up() can work "inside" ptrace_request()
even if the tracee doesn't have the TASK_WAKEKILL bit set.
Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Colin King <colin.king@canonical.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
Oleg Nesterov [Tue, 19 Feb 2013 17:31:07 +0000 (17:31 +0000)]
ptrace: ptrace_resume() shouldn't wake up !TASK_TRACED thread
ptrace: ptrace_resume() shouldn't wake up !TASK_TRACED thread
CVE-2013-0871
BugLink: http://bugs.launchpad.net/bugs/1129192
It is not clear why ptrace_resume() does wake_up_process(). Unless the
caller is PTRACE_KILL the tracee should be TASK_TRACED so we can use
wake_up_state(__TASK_TRACED). If sys_ptrace() races with SIGKILL we do
not need the extra and potentionally spurious wakeup.
If the caller is PTRACE_KILL, wake_up_process() is even more wrong.
The tracee can sleep in any state in any place, and if we have a buggy
code which doesn't handle a spurious wakeup correctly PTRACE_KILL can
be used to exploit it. For example:
prints "pause: -1 Unknown error 514", -ERESTARTNOHAND leaks to the
userland. In this case sys_pause() is buggy as well and should be
fixed.
I do not know what was the original rationality behind PTRACE_KILL.
The man page is simply wrong and afaics it was always wrong. Imho
it should be deprecated, or may be it should do send_sig(SIGKILL)
as Denys suggests, but in any case I do not think that the current
behaviour was intentional.
Note: there is another problem, ptrace_resume() changes ->exit_code
and this can race with SIGKILL too. Eventually we should change ptrace
to not use ->exit_code.
Signed-off-by: Luis Henriques <luis.henriques@canonical.com> Acked-by: Colin King <colin.king@canonical.com> Signed-off-by: Tim Gardner <tim.gardner@canonical.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
When the new signal handlers are set up, the location of sa_restorer is
not cleared, leaking a parent process's address space location to
children. This allows for a potential bypass of the parent's ASLR by
examining the sa_restorer value returned when calling sigaction().
Based on what should be considered "secret" about addresses, it only
matters across the exec not the fork (since the VMAs haven't changed
until the exec). But since exec sets SIG_DFL and keeps sa_restorer,
this is where it should be fixed.
Given the few uses of sa_restorer, a "set" function was not written
since this would be the only use. Instead, we use
__ARCH_HAS_SA_RESTORER, as already done in other places.
To avoid an explosion of request_module calls on a chain of abusive
scripts, fail maximum recursion with -ELOOP instead of -ENOEXEC. As soon
as maximum recursion depth is hit, the error will fail all the way back
up the chain, aborting immediately.
This also has the side-effect of stopping the user's shell from attempting
to reexecute the top-level file as a shell script. As seen in the
dash source:
The above logic was designed for running scripts automatically that lacked
the "#!" header, not to re-try failed recursion. On a legitimate -ENOEXEC,
things continue to behave as the shell expects.
Additionally, when tracking recursion, the binfmt handlers should not be
involved. The recursion being tracked is the depth of calls through
search_binary_handler(), so that function should be exclusively responsible
for tracking the depth.
Signed-off-by: Kees Cook <keescook@chromium.org> Cc: halfdog <me@halfdog.net> Cc: P J P <ppandit@redhat.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[dannf: backported to Debian's 2.6.32] Signed-off-by: Willy Tarreau <w@1wt.eu>
If a series of scripts are executed, each triggering module loading via
unprintable bytes in the script header, kernel stack contents can leak
into the command line.
Normally execution of binfmt_script and binfmt_misc happens recursively.
However, when modules are enabled, and unprintable bytes exist in the
bprm->buf, execution will restart after attempting to load matching
binfmt modules. Unfortunately, the logic in binfmt_script and
binfmt_misc does not expect to get restarted. They leave bprm->interp
pointing to their local stack. This means on restart bprm->interp is
left pointing into unused stack memory which can then be copied into the
userspace argv areas.
After additional study, it seems that both recursion and restart remains
the desirable way to handle exec with scripts, misc, and modules. As
such, we need to protect the changes to interp.
This changes the logic to require allocation for any changes to the
bprm->interp. To avoid adding a new kmalloc to every exec, the default
value is left as-is. Only when passing through binfmt_script or
binfmt_misc does an allocation take place.
As Tetsuo Handa pointed out, request_module() can stress the system
while the oom-killed caller sleeps in TASK_UNINTERRUPTIBLE.
The task T uses "almost all" memory, then it does something which
triggers request_module(). Say, it can simply call sys_socket(). This
in turn needs more memory and leads to OOM. oom-killer correctly
chooses T and kills it, but this can't help because it sleeps in
TASK_UNINTERRUPTIBLE and after that oom-killer becomes "disabled" by the
TIF_MEMDIE task T.
Make __request_module() killable. The only necessary change is that
call_modprobe() should kmalloc argv and module_name, they can't live in
the stack if we use UMH_KILLABLE. This memory is freed via
call_usermodehelper_freeinfo()->cleanup.
Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Signed-off-by: Oleg Nesterov <oleg@redhat.com> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[dannf, bwh: backported to Debian's 2.6.32] Signed-off-by: Willy Tarreau <w@1wt.eu>
Implement UMH_KILLABLE, should be used along with UMH_WAIT_EXEC/PROC.
The caller must ensure that subprocess_info->path/etc can not go away
until call_usermodehelper_freeinfo().
call_usermodehelper_exec(UMH_KILLABLE) does
wait_for_completion_killable. If it fails, it uses
xchg(&sub_info->complete, NULL) to serialize with umh_complete() which
does the same xhcg() to access sub_info->complete.
If call_usermodehelper_exec wins, it can safely return. umh_complete()
should get NULL and call call_usermodehelper_freeinfo().
Otherwise we know that umh_complete() was already called, in this case
call_usermodehelper_exec() falls back to wait_for_completion() which
should succeed "very soon".
Note: UMH_NO_WAIT == -1 but it obviously should not be used with
UMH_KILLABLE. We delay the neccessary cleanup to simplify the back
porting.
Signed-off-by: Oleg Nesterov <oleg@redhat.com> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Rusty Russell <rusty@rustcorp.com.au> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[dannf: backported to Debian's 2.6.32] Signed-off-by: Willy Tarreau <w@1wt.eu>
Fix possible overflow of the buffer used for expanding environment
variables when building file list.
In the extremely unlikely case of an attacker having control over the
environment variables visible to gen_init_cpio, control over the
contents of the file gen_init_cpio parses, and gen_init_cpio was built
without compiler hardening, the attacker can gain arbitrary execution
control via a stack buffer overflow.
The correct syntax for gcc -x is "gcc -x assembler", not
"gcc -xassembler". Even though the latter happens to work, the former
is what is documented in the manual page and thus what gcc wrappers
such as icecream do expect.
This isn't a cosmetic change. The missing space prevents icecream from
recognizing compilation tasks it can't handle, leading to silent kernel
miscompilations.
Besides me, credits go to Michael Matz and Dirk Mueller for
investigating the miscompilation issue and tracking it down to this
incorrect -x parameter syntax.
Signed-off-by: Jean Delvare <jdelvare@suse.de> Cc: stable@vger.kernel.org Cc: Bernhard Walle <bernhard@bwalle.de> Cc: Michal Marek <mmarek@suse.cz> Cc: Ralf Baechle <ralf@linux-mips.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
Prarit reported a crash on CPU offline/online. The reason is that on
CPU down the NOHZ related per cpu data of the dead cpu is not cleaned
up. If at cpu online an interrupt happens before the per cpu tick
device is registered the irq_enter() check potentially sees stale data
and dereferences a NULL pointer.
An inactive timer's base can refer to a offline cpu's base.
In the current code, cpu_base's lock is blindly reinitialized each
time a CPU is brought up. If a CPU is brought online during the period
that another thread is trying to modify an inactive timer on that CPU
with holding its timer base lock, then the lock will be reinitialized
under its feet. This leads to following SPIN_BUG().
<0> BUG: spinlock already unlocked on CPU#3, kworker/u:3/1466
<0> lock: 0xe3ebe000, .magic: dead4ead, .owner: kworker/u:3/1466, .owner_cpu: 1
<4> [<c0013dc4>] (unwind_backtrace+0x0/0x11c) from [<c026e794>] (do_raw_spin_unlock+0x40/0xcc)
<4> [<c026e794>] (do_raw_spin_unlock+0x40/0xcc) from [<c076c160>] (_raw_spin_unlock+0x8/0x30)
<4> [<c076c160>] (_raw_spin_unlock+0x8/0x30) from [<c009b858>] (mod_timer+0x294/0x310)
<4> [<c009b858>] (mod_timer+0x294/0x310) from [<c00a5e04>] (queue_delayed_work_on+0x104/0x120)
<4> [<c00a5e04>] (queue_delayed_work_on+0x104/0x120) from [<c04eae00>] (sdhci_msm_bus_voting+0x88/0x9c)
<4> [<c04eae00>] (sdhci_msm_bus_voting+0x88/0x9c) from [<c04d8780>] (sdhci_disable+0x40/0x48)
<4> [<c04d8780>] (sdhci_disable+0x40/0x48) from [<c04bf300>] (mmc_release_host+0x4c/0xb0)
<4> [<c04bf300>] (mmc_release_host+0x4c/0xb0) from [<c04c7aac>] (mmc_sd_detect+0x90/0xfc)
<4> [<c04c7aac>] (mmc_sd_detect+0x90/0xfc) from [<c04c2504>] (mmc_rescan+0x7c/0x2c4)
<4> [<c04c2504>] (mmc_rescan+0x7c/0x2c4) from [<c00a6a7c>] (process_one_work+0x27c/0x484)
<4> [<c00a6a7c>] (process_one_work+0x27c/0x484) from [<c00a6e94>] (worker_thread+0x210/0x3b0)
<4> [<c00a6e94>] (worker_thread+0x210/0x3b0) from [<c00aad9c>] (kthread+0x80/0x8c)
<4> [<c00aad9c>] (kthread+0x80/0x8c) from [<c000ea80>] (kernel_thread_exit+0x0/0x8)
As an example, this particular crash occurred when CPU #3 is executing
mod_timer() on an inactive timer whose base is refered to offlined CPU
#2. The code locked the timer_base corresponding to CPU #2. Before it
could proceed, CPU #2 came online and reinitialized the spinlock
corresponding to its base. Thus now CPU #3 held a lock which was
reinitialized. When CPU #3 finally ended up unlocking the old cpu_base
corresponding to CPU #2, we hit the above SPIN_BUG().
CPU #0 CPU #3 CPU #2
------ ------- -------
..... ...... <Offline>
mod_timer()
lock_timer_base
spin_lock_irqsave(&base->lock)
Allocation of per_cpu timer vector bases is done only once under
"tvec_base_done[]" check. In the current code, spinlock_initialization
of base->lock isn't under this check. When a CPU is up each time the
base lock is reinitialized. Move base spinlock initialization under
the check.
The trinity fuzzer triggered a task_struct reference leak via
clock_nanosleep with CPU_TIMERs. do_cpu_nanosleep() calls
posic_cpu_timer_create(), but misses a corresponding
posix_cpu_timer_del() which leads to the task_struct reference leak.
Reported-and-tested-by: Tommi Rantala <tt.rantala@gmail.com> Signed-off-by: Stanislaw Gruszka <sgruszka@redhat.com> Cc: Dave Jones <davej@redhat.com> Cc: John Stultz <john.stultz@linaro.org> Cc: Oleg Nesterov <oleg@redhat.com> Link: http://lkml.kernel.org/r/20130215100810.GF4392@redhat.com Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
Currently tick_check_broadcast_device doesn't reject clock_event_devices
with CLOCK_EVT_FEAT_DUMMY, and may select them in preference to real
hardware if they have a higher rating value. In this situation, the
dummy timer is responsible for broadcasting to itself, and the core
clockevents code may attempt to call non-existent callbacks for
programming the dummy, eventually leading to a panic.
This patch makes tick_check_broadcast_device always reject dummy timers,
preventing this problem.
Signed-off-by: Mark Rutland <mark.rutland@arm.com> Cc: linux-arm-kernel@lists.infradead.org Cc: Jon Medhurst (Tixy) <tixy@linaro.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Willy Tarreau <w@1wt.eu>
Basically with kernels between 2.6.20-something to 2.6.32,
we accumulate time in half second chunks, rather then every
timer-tick. This was added because when NOHZ landed, if you
were idle for a few seconds, you had to spin for every tick
we skipped in the accumulation loop, which created some bad
latencies.
However, this required that we create the xtime_cache() which
was still updated each tick, so that filesystem timestamps,
etc continued to see time increment normally.
Of course, the xtime_cache is updated at the bottom of
update_wall_time(). So the early return on
(offset < timekeeper.cycle_interval), added by the problematic
commit causes the xtime_cache to not be updated.
This can cause code using current_kernel_time() (like the mqueue
code) or hrtimer_get_softirq_time(), which uses the non-updated
xtime_cache, to see timers to fire with very coarse half-second
granularity.
Many thanks to Romain for describing the issue clearly,
providing test case to reproduce it and helping with testing
the solution.
"Benjamin S." <sbenni@gmx.de> reports that the patch in question
causes a big drop in sequential throughput for him, dropping from
200MB/sec down to only 70MB/sec.
Needs to be investigated more fully, for now lets just revert the
offending commit.
Conflicts:
include/linux/blkdev.h
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
(cherry picked from commit 79da0644a8e0838522828f106e4049639eea6baf) Cc: Thomas Bork <tom@eisfair.net> Signed-off-by: Willy Tarreau <w@1wt.eu>
This broke compilation of the driver in 2.6.32.y as the
early_io{remap,unmap}() functions are not defined for ia64. The driver
can *only* be built for ia64 (even in current mainline), so a fix for
x86_64 is pointless.
Signed-off-by: Ben Hutchings <ben@decadent.org.uk> Signed-off-by: Willy Tarreau <w@1wt.eu>
The tamper evident features of the RTC include the "write counter" which
is a pseudo-random number regenerated whenever we set the RTC. Since this
value is unpredictable it should provide some useful seeding to the random
number generator.
Only do this on boot since the goal is to seed the pool rather than add
useful entropy.
Signed-off-by: Mark Brown <broonie@opensource.wolfsonmicro.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
Many platforms have per-machine instance data (serial numbers,
asset tags, etc.) squirreled away in areas that are accessed
during early system bringup. Mixing this data into the random
pools has a very high value in providing better random data,
so we should allow (and even encourage) architecture code to
call add_device_randomness() from the setup_arch() paths.
However, this limits our options for internal structure of
the random driver since random_initialize() is not called
until long after setup_arch().
Add a big fat comment to rand_initialize() spelling out
this requirement.
Suggested-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Tony Luck <tony.luck@intel.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
With the new interrupt sampling system, we are no longer using the
timer_rand_state structure in the irq descriptor, so we can stop
initializing it now.
[ Merged in fixes from Sedat to find some last missing references to
rand_initialize_irq() ]
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu> Signed-off-by: Sedat Dilek <sedat.dilek@gmail.com>
[PG: in .34 the irqdesc.h content is in irq.h instead.] Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
Mix in any architectural randomness in extract_buf() instead of
xfer_secondary_buf(). This allows us to mix in more architectural
randomness, and it also makes xfer_secondary_buf() faster, moving a
tiny bit of additional CPU overhead to process which is extracting the
randomness.
[ Commit description modified by tytso to remove an extended
advertisement for the RDRAND instruction. ]
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Acked-by: Ingo Molnar <mingo@kernel.org> Cc: DJ Johnston <dj.johnston@intel.com> Signed-off-by: Theodore Ts'o <tytso@mit.edu> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
Create a new function, get_random_bytes_arch() which will use the
architecture-specific hardware random number generator if it is
present. Change get_random_bytes() to not use the HW RNG, even if it
is avaiable.
The reason for this is that the hw random number generator is fast (if
it is present), but it requires that we trust the hardware
manufacturer to have not put in a back door. (For example, an
increasing counter encrypted by an AES key known to the NSA.)
It's unlikely that Intel (for example) was paid off by the US
Government to do this, but it's impossible for them to prove otherwise
--- especially since Bull Mountain is documented to use AES as a
whitener. Hence, the output of an evil, trojan-horse version of
RDRAND is statistically indistinguishable from an RDRAND implemented
to the specifications claimed by Intel. Short of using a tunnelling
electronic microscope to reverse engineer an Ivy Bridge chip and
disassembling and analyzing the CPU microcode, there's no way for us
to tell for sure.
Since users of get_random_bytes() in the Linux kernel need to be able
to support hardware systems where the HW RNG is not present, most
time-sensitive users of this interface have already created their own
cryptographic RNG interface which uses get_random_bytes() as a seed.
So it's much better to use the HW RNG to improve the existing random
number generator, by mixing in any entropy returned by the HW RNG into
/dev/random's entropy pool, but to always _use_ /dev/random's entropy
pool.
This way we get almost of the benefits of the HW RNG without any
potential liabilities. The only benefits we forgo is the
speed/performance enhancements --- and generic kernel code can't
depend on depend on get_random_bytes() having the speed of a HW RNG
anyway.
For those places that really want access to the arch-specific HW RNG,
if it is available, we provide get_random_bytes_arch().
If the CPU supports a hardware random number generator, use it in
xfer_secondary_pool(), where it will significantly improve things and
where we can afford it.
Also, remove the use of the arch-specific rng in
add_timer_randomness(), since the call is significantly slower than
get_cycles(), and we're much better off using it in
xfer_secondary_pool() anyway.
Add a new interface, add_device_randomness() for adding data to the
random pool that is likely to differ between two devices (or possibly
even per boot). This would be things like MAC addresses or serial
numbers, or the read-out of the RTC. This does *not* add any actual
entropy to the pool, but it initializes the pool to different values
for devices that might otherwise be identical and have very little
entropy available to them (particularly common in the embedded world).
[ Modified by tytso to mix in a timestamp, since there may be some
variability caused by the time needed to detect/configure the hardware
in question. ]
The real-time Linux folks don't like add_interrupt_randomness() taking
a spinlock since it is called in the low-level interrupt routine.
This also allows us to reduce the overhead in the fast path, for the
random driver, which is the interrupt collection path.
We've been moving away from add_interrupt_randomness() for various
reasons: it's too expensive to do on every interrupt, and flooding the
CPU with interrupts could theoretically cause bogus floods of entropy
from a somewhat externally controllable source.
This solves both problems by limiting the actual randomness addition
to just once a second or after 64 interrupts, whicever comes first.
During that time, the interrupt cycle data is buffered up in a per-cpu
pool. Also, we make sure the the nonblocking pool used by urandom is
initialized before we start feeding the normal input pool. This
assures that /dev/urandom is returning unpredictable data as soon as
possible.
(Based on an original patch by Linus, but significantly modified by
tytso.)
Tested-by: Eric Wustrow <ewust@umich.edu> Reported-by: Eric Wustrow <ewust@umich.edu> Reported-by: Nadia Heninger <nadiah@cs.ucsd.edu> Reported-by: Zakir Durumeric <zakir@umich.edu> Reported-by: J. Alex Halderman <jhalderm@umich.edu>. Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
[PG: minor adjustment required since .34 doesn't have f9e4989eb8
which renames "status" to "random" in kernel/irq/handle.c ] Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
/proc/sys/kernel/random/boot_id can be read concurrently by userspace
processes. If two (or more) user-space processes concurrently read
boot_id when sysctl_bootid is not yet assigned, a race can occur making
boot_id differ between the reads. Because the whole point of the boot id
is to be unique across a kernel execution, fix this by protecting this
operation with a spinlock.
Given that this operation is not frequently used, hitting the spinlock
on each call should not be an issue.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com> Cc: "Theodore Ts'o" <tytso@mit.edu> Cc: Matt Mackall <mpm@selenic.com> Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> Cc: Greg Kroah-Hartman <greg@kroah.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
When we are initializing using arch_get_random_long() we only need to
loop enough times to touch all the bytes in the buffer; using
poolwords for that does twice the number of operations necessary on a
64-bit machine, since in the random number generator code "word" means
32 bits.
If there is an architecture-specific random number generator (such as
RDRAND for Intel architectures), use it to initialize /dev/random's
entropy stores. Even in the worst case, if RDRAND is something like
AES(NSA_KEY, counter++), it won't hurt, and it will definitely help
against any other adversaries.
We still don't use rdrand in /dev/random, which just seems stupid. We
accept the *cycle*counter* as a random input, but we don't accept
rdrand? That's just broken.
Sure, people can do things in user space (write to /dev/random, use
rdrand in addition to /dev/random themselves etc etc), but that
*still* seems to be a particularly stupid reason for saying "we
shouldn't bother to try to do better in /dev/random".
And even if somebody really doesn't trust rdrand as a source of random
bytes, it seems singularly stupid to trust the cycle counter *more*.
So I'd suggest the attached patch. I'm not going to even bother
arguing that we should add more bits to the entropy estimate, because
that's not the point - I don't care if /dev/random fills up slowly or
not, I think it's just stupid to not use the bits we can get from
rdrand and mix them into the strong randomness pool.
Link: http://lkml.kernel.org/r/CA%2B55aFwn59N1=m651QAyTy-1gO1noGbK18zwKDwvwqnravA84A@mail.gmail.com Acked-by: "David S. Miller" <davem@davemloft.net> Acked-by: "Theodore Ts'o" <tytso@mit.edu> Acked-by: Herbert Xu <herbert@gondor.apana.org.au> Cc: Matt Mackall <mpm@selenic.com> Cc: Tony Luck <tony.luck@intel.com> Cc: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
If there is an architecture-specific random number generator we use it
to acquire randomness one "long" at a time. We should put these random
words into consecutive words in the result buffer - not just overwrite
the first word again and again.
Signed-off-by: Tony Luck <tony.luck@intel.com> Acked-by: H. Peter Anvin <hpa@zytor.com> Acked-by: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
If the CPU declares that RDRAND is available, go through a guranteed
reseed sequence, and make sure that it is actually working (producing
data.) If it does not, disable the CPU feature flag.
Allow RDRAND to be disabled on the command line (as opposed to at
compile time) for a user who has special requirements with regards to
random numbers.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Cc: Matt Mackall <mpm@selenic.com> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: "Theodore Ts'o" <tytso@mit.edu> Signed-off-by: Willy Tarreau <w@1wt.eu>
Architectural inlines to get random ints and longs using the RDRAND
instruction.
Intel has introduced a new RDRAND instruction, a Digital Random Number
Generator (DRNG), which is functionally an high bandwidth entropy
source, cryptographic whitener, and integrity monitor all built into
hardware. This enables RDRAND to be used directly, bypassing the
kernel random number pool.
In this patch, this is *only* used for the nonblocking random number
pool. RDRAND is a nonblocking source, similar to our /dev/urandom,
and is therefore not a direct replacement for /dev/random. The
architectural hooks presented in the previous patch only feed the
kernel internal users, which only use the nonblocking pool, and so
this is not a problem.
Since this instruction is available in userspace, there is no reason
to have a /dev/hw_rng device driver for the purpose of feeding rngd.
This is especially so since RDRAND is a nonblocking source, and needs
additional whitening and reduction (see the above technical
documentation for details) in order to be of "pure entropy source"
quality.
The CONFIG_EXPERT compile-time option can be used to disable this use
of RDRAND.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> Originally-by: Fenghua Yu <fenghua.yu@intel.com> Cc: Matt Mackall <mpm@selenic.com> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: "Theodore Ts'o" <tytso@mit.edu> Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Willy Tarreau <w@1wt.eu>
projects / people / ms / linux.git / log
